Proceedings of Student/Faculty Research Day, CSIS, Pace University, May 3, 2013
A New Algorithmic Approach to the IT Modernization Problem
David Fronckowiak and Robert Zandoli
Seidenberg School of CSIS, Pace University, White Plains, NY 10606
Abstract
This paper contains the research plan to develop a new simple, parsimonious algorithm to assist in the IT modernization decision making process. This algorithm will be developed based upon known, well-established models that exist in the area of
decision making and technology
acceptance. Introduction
With the latest IT technology advancements in analytics, big data, Saas, cloud computing, social networks and mobile computing, companies are faced with the decision of “if and how” these technologies should be implemented In 2005, Zach Nelson and Joe Gentry [1] presented their views on whether a company should replace their legacy system or extend the lifetime of their legacy system. Nelson makes the case that companies should replace their legacy systems to remain competitive, comply with regulatory issues, and improve usability. Gentry, however, is an advocate for the “preserve and extend” philosophy whereby the system should continue to function as designed and implemented or be extended through the addition of “minor enhancements.” Seacord et al. [2] dedicate the first chapter in their book to “The Legacy Crisis.” The crisis
relates to the concern that “the development of new software is outpacing our ability to maintain it.” With the continued maintenance of these legacy systems, no resources will be left to develop new software systems. So in terms of the decision process, does the fact that a system is considered legacy change the process which is used to manage the system or is the same decision process used for all IT investment?
Numerous articles have been written on IT modernization, but few have focused on the decision process and the key drivers that are critical to the extension or replacement decision in light of the entire system. The literature can be divided into those works which present strategies that can be used in replacing or extending the lifetime of a legacy system, technologies that are applicable, legacy system assessment frameworks, and legacy system management tools which also include the legacy system extension or replacement decision process.
In terms of strategies which can be used to replace or extend the lifetime of a legacy system, Lavelle [3] mentions wait, wrap, renovate, replace, and outsource. If a product or business line is nearing end-of-life, then the wait strategy is ideal for no investment or decision is required.
Wrapping a legacy system in middleware may enhance functionality, but increase complexity and maintenance costs. Renovate refers to the conversion of a legacy system into a more modern form. Lavelle uses the example of renovating a COBOL system into J2EE technology. Replacement is another strategy in which the current legacy system is replaced with newer technology, but this effort may be expensive and require several years given the complexity of the system. Finally, an outsource strategy may be incorporated where the business focuses on its core competence and outsources its IT needs to a company which specializes in that service. With the advent of the previously mentioned technologies, namely, Saas, analytics, social networking, cloud computing and mobile computing, new issues arise in the IT modernization dilemma. Wand and Quin [4] described component-oriented programming. With a component-oriented approach, the system can be subdivided into components and then based upon functionality, risk, skill, and cost constraints, a component can be highlighted as to one which needs to be enhanced or replaced. The problem with this technique is that the entire system must be taken into consideration in this component modernization and a simple algorithm to recommend which component or subsystem should be modernized would be ideal.
Examples of IT Modernization Problems
Over the past several years, three companies stand out with respect to the impact that IT modernization or the lack of it can have on the company’s financial statements. HP’s 1Q 2012 earnings were down 44% and one reason cited in the quarterly financial report for a portion of
the decline was that HP “didn't make the
investments they should have during the past few years to stay ahead of customer expectations and market trends. As a result, they saw eroding revenue and profits. They need to invest now as a market leader from a position of strength, and that's especially true because these businesses are not only under intense competitive pressure but are also under pressure from tectonic shifts that are taking place at the very foundation of the industry.”[5] This emphasizes the potential impact that IT modernization has on a company’s revenue stream and how it must be considered in the overall business plan. Investment with respect to cost and affordability must be considered in the IT modernization decision making process. Along with cost, risk is a key component that must be addressed. In June of 2012, the Royal bank of Scotland (RBS) experienced an ATM outage for five days which was hinted to be caused by an issue with an IT upgrade. [6] In October of 2009, TD Bank experienced a glitch related to its merger with Commerce bank leaving customers unable to access funds. [7] This risk involved in IT upgrades and IT modernization is a key factor in planning and executing an IT modernization plan.
IT Modernization Decision Process
Problem
This proposed study will focus on the IT modernization decision making process used in the extension, replacement or enhancement of IT systems. Known, well-accepted models will be modified to develop a framework and algorithm to assist in the IT modernization management decision process. Figure 1, which shows an example system architecture view, can be used to illustrate the research problem in question. With the many component and subsystem interrelationships depicted, it
may be very difficult to ascertain which component or subsystem should be modernized based upon needed enhanced functionality, cost and risk, to name a few driving factors.
Models that are appropriate for IT System Modernization Management
One well accepted model is applicable in the IT system extension, replacement, or enhancement decision making process. Simon’s decision model [8], as shown in figure 2, can be used to model the overall
decision process.
Figure 1. System Architecture Example
Figure 2. Simon’s Descriptive Decision Model Intelligence
Identify need for a decision Design Detail problem domain & alternative solutions Choice Select appropriate actions Implementation Webpublishing Subsystem Component A Component B Sub Component C Component D Subsystem Subsystem Subsystem 1 Component F Component G Component H Subsystem Component E Subsystem 1 Subsystem 2 Subsystem 3 Subsystem 2 Subsystem 1 Subsystem 1 Subsystem 2 Subsystem 3 Transaction volumes in italics 10,000 30,000 10,000 45,000 15,000 5,000 5,000 20,000 3,000 98% reliability 98% reliability 96% reliability 99% reliability 95% reliability
This model can be modified to show the drivers or factors which influence the blocks in the model and the links between each of the blocks in the model. In Simon’s decision model, the links indicate an order in which the decision process occurs along
with an iterative factor that exists when new information becomes relevant. With this model, the drivers, as shown in figure 3, are critical in the development of the model. Table 1 lists potential modernization drivers.
Figure 3. Simon’s Descriptive Decision Model with Relationship Factors
Business or technical driver Decision Driver
Business Operations costs Business Maintenance/Support costs Business Customer access /ability Business Legal requirements Technical Web Services Business Scheduled need date Technical Virtualization Business / Technical User Mobility Business / Technical COTS
Business Risk
Technical Internal development skills Business Benchmarking
Technical Component redesign
Technical SOA
Decision process Technical champion Decision process Business champion Decision process Budget
Decision process Review board
Intelligence
Identify need for a decision Design Detail problem domain & alternative solutions Choice Select appropriate actions Implementation
Driver 2 Driver 3 Driverr 4 Driver 5 Driver 1
Table 1. IT Modernization Management Decision Drivers
Figure 4 shows the drivers as applied to Simon’s decision making model.
Figure 4. Legacy management decision drivers and Simon’s decision making model
Research Methodology
To analyze the IT modernization decision making process and develop an algorithm to assist in this decision making process the key drivers must be identified through a survey of experts. Previous work shows these drivers in table 1. Once the key drivers are identified, with a quantifiable metric, an algorithm can be developed using these drivers to highlight the critical component or subsystem recommended for modernization in light of the entire system. It may be suggested to use qualitative surveys based upon benefits, rate of return, complexity, and time to adoption, but often this qualitative information and benefit may be overstated. Quantifiable factors such as
inter-component transactions, customer satisfaction, and failure rates are examples of key quantifiable drivers which can be incorporated into an index style algorithm which can be used to identify the component or subsystem which should be modernized to improve overall system, functionality, performance, and reliability. Decision making models and algorithms such as critical path, network flow algorithms, linear programming, integer programming, and dynamic programming [9] are all potential decision making techniques, but their complexity as related to the decision recommendation is difficult to explain. A simple, easy-to-explain from-to analysis based upon the quantifiable critical business factors such as
Intelligence
Identify need for a decision Design Detail problem domain & alternative solutions
Choice Select appropriate actions Legal Review board Maint / Support Customer access / usability COTS
Web Services Sat / Performance Metrics
User Mobility Technical Champion Business Champion Virtualization Component redesign Executive Sponsor Risk Need Date Extend Replace Imp le men tat ion Operations Cost Benchmarking Budget
maintenance cost, component and subsystem failure rates, and transaction volumes will be incorporated into this type of IT modernization decision making algorithm.
The study will focus on the following topics with respect to the IT modernization problem in light of the entire system:
(1) What are the key drivers in the IT modernization management decision making process?
(2) Can these drivers be mapped to an existing decision making model? (3) Develop an index based upon
quantifiable IT modernization factors.
(4) Develop a from/to style algorithm to recommend IT modernization for components and subsystems.
(5) Test the algorithm through expert test.
Future Work
From / to algorithms can be developed and tested through the use of spreadsheets. An interesting extension would be to develop the algorithm using the statistical package R from the Foundation for Statistical Computing software which has become very well accepted in the statistical and decision making environments.. This algorithmic index can be tested across different industries that may have unique IT modernization characteristics which relate to the entire system under analysis.
References
1. Nelson, Z. and J. Gentry, Face-off: Should your company replace its legacy systems?, in
Network World. April 25, 2005. p. 49.
2. Seacord, R.C., D. Plakosh, and G.A. Lewis, Modernizing Legacy Systems. 2003, Boston, Ma.: Addison-Wesley.
3. Lavelle, G., Practical Strategies for Dealing with Legacy Systems. Risk Management, Feb 2005. 52(2): p. 45.
4. Wang, A. and K. Qian, Component Oriented Programming. 2005, Hoboken, N.J.: John Wiley & Sons.
5. Hewlett-Packard 2012 Quarterly Financial Reports,
http://h30261.www3.hp.com/phoenix.zhtml?c=71087&p=quarterlyEarnings.
6. Durden, Tyler (June 23, 2012) "As RBS' ATM "Glitch" Enters Fifth Day, The Bailed out bank Issues a Statement, http://www.zerohedge.com/news/rbs-atm-glitch-enters-fifth-day-bailed-out-bank-issues-statement.
7. Customers Up in Arms over TD Bank Glitches, (Oct. 2, 2009)
http://www.nbcphiladelphia.com/news/business/Computer-Glitch-Causes-Problems-at-TD-Bank-63103572.html.
8. Simon, H.A., The New Science of Management Decision. 1960, New York, New York: Harper and Row. 50.
9. Gass, S.I., Desision Making, Models and Algorithms. 1985, Malabar, Florida: Krieger Publishing Company.
